Global Land Use Changes due to US Cellulosic Biofuel Program: A Preliminary Analysis And Updated Corn Ethanol, Biodiesel, and Sugarcane Ethanol Estimates Wallace E. Tyner, Professor In collaboration with Farzad Taheripour Purdue University Michael Wang Argonne National Lab October 18, 2011 Argonne National Lab 1
Steps in Our Analysis Introduce the first generation if biofuels into version 7 of the GTAP data base (2004). All the prior work was done on ver. 6 (2001). Introduce new cellulosic biofuels and their supporting activities into the model. Add greater flexibility in acreage switching among crops in the US in response to price changes. Include an endogenous yield adjustment for cropland pasture in response to changes in cropland pasture rent. 2
New GTAP-BIO Database Introduced 2004 global production, consumption, and trade for first generation biofuels. Updated land use, land cover, and land rent headers to 2004. Following the previous work, created new industries for corn ethanol, sugarcane ethanol, and biodiesel. Modified the basic GTAP database as was done previously: Split GTAP food industry into food and feed industries, Split GTAP vegetable oil into crude and refined vegetable oil industries. Introduced by-products into the 2004 database. 3
Introduced Cellulosic Feedstock and Biofuels Industries into Version 7 Corn stover industry which collects corn stover from corn land and delivers it to the cellulosic biofuel industry. Dedicated crop industries (miscanthus and switchgrass) produce feedstock and deliver it to biofuel industries. Introduced a biofuel (bio-gasoline) processing industry and ethanol for each feedstock with identical cost structures. Since none of these industries exist, we developed consensus estimates using experts from Argonne, NREL, and Purdue for dedicated energy crop yields and conversion technologies. 4
Biofuel Production Costs Cost Items FeedStock ($ / dry short ton) Thermo - Gasoline Pathways Bio - Ethanol - Stover Bio - Ethanol - Dedicated Crops Capital cost ($/gal.) $1.14 $0.51 $0.57 Operating cost ($/gal.) $0.49 $1.34 $1.52 Feedstock cost: Stover ($/gal) $89.47 $1.49 $1.19 Switchgrass ($/gal) $121.37 $2.02 $1.62 Miscanthus ($/gal) $126.03 $2.10 $1.68 Total cost - stover $3.12 $3.05 Total cost - switchgrass $3.65 $3.71 Total cost - miscanthus $3.73 $3.77 Assumed that the conversion rate is 60 gallons of biogasoline per metric ton of feedstock and 75 gallons of ethanol per dry ton. 5
Cost Structures for Feedstocks Cost Items Corn Stover Miscanthus Switchgrass Fertilizer 22.7 14.0 15.6 Transportation 33.5 25.4 28.4 Fuel 3.4 4.6 5.1 Payments to seed company 0.0 6.7 1.7 Other costs 7.0 7.5 8.0 Labor 10.0 10.7 11.5 Land 0.0 2.7 5.8 Capital (including profit) 23.3 28.5 23.9 Total 100.0 100.0 100.0 6
Cost Structures for Corn Stover, Miscanthus, and Switchgrass Cost items Miscanthus Bio-gasoline Switchgrass Corn stover Ethanol Miscanthus Switchgrass Corn stover Feedstock 54.6 51.9 47.7 42.9 40.2 39.2 Chemicals 0.0 0.0 0.0 15.6 16.3 18.8 Energy 1.0 1.0 1.1 4.1 4.2 4.9 Other costs 10.5 11.1 12.1 17.5 18.3 15.0 Labor 2.2 2.4 2.6 4.4 4.6 5.3 Capital 31.8 33.7 36.6 15.6 16.3 16.9 Total 100.0 100.0 100.0 100.0 100.0 100.0 7
substitute for fossil fuels and biofuels. Figures B-1 and B-2 represent these demands. Figure Household Demand Structure in the GTAP-BIO- ADVFUEL Model Household Demand for Private Goods CDE Energy Composite Non-Energy Commodities CES s ELEGY Coal Oil Gas Electricity Petroleum products & Biofuels composite s ELHBIOIL Petro products (Oil_Pcts) Biodiesel Biofuels s biofuels Ethanol Bio-gasoline s ethanol s bio-gasoline Eth1 Eth2 Eth_Misc Eth_Swit Eth_Stover Biog_Misc Biog_Swit Biog_Stover 8
Transformation elasticity among cropland pasture, miscanthus, and switchgrass = - 10 Old Model Land Supply Forest Pasture Cropland Land Cover and Land Use Activities in the GTAP-BIO- ADVFUEL Model Cropland- Pasture New Model CRP Crop 1 Crop n Land Supply 1 Fore st Pastur e Cropland 2 Crop group 2 Crop group 1 Cropland- Pasture 5 Miscanthus Switchgrass CRP 4 Crop 1 Crop N 9
Add Greater Flexibility in Acreage Switching Among Crops In our previous work we and others had observed that GTAP does not seem to have as much acreage responsiveness as we experienced in the decade 2000-09. In this analysis, we asked the question of whether there is any difference in farmers reactions to crop price changes in the past decade and earlier periods. 10
Add Greater Flexibility in Acreage Switching Among Crops To answer this question we estimated acreage response to changes in soybean and corn returns per acre over different decades prior to 2000 and for 2000-2010. The following regression shows the results for the time period of 2000-2010: Harvested corn area (acres) = 1.388 + 0.084 Corn revenue/acre(t-1) 0.138 Soybean revenue/acre(t-1), The independent variable t values are 2.9 and 3.0 respectively, and the adjusted R 2 is 0.44. We did the same regressions for prior periods and found no significant relationship. 11
Add Greater Flexibility in Acreage Switching Among Crops As the literature suggests, in prior periods, government policy was a major driver, and now it is commodity prices and revenue. For these reasons, we increased the transformation elasticity that helps govern the response in acreage share to changes in commodity prices from 0.5 to 0.75. However, we are still experimenting with this parameter value to make sure it is the best representation of reality possible. 12
Endogenous Cropland Pasture Yield Change We received comments on our previous work suggesting that the increased use of land for biofuels would lead to investments in increased productivity as land rents increased. This led us to introduce an endogenous change in cropland pasture productivity as cropland pasture rent increases due to higher demand for the resource. This change in productivity is a function of the change in rent and a new elasticity parameter. 13
Scenarios Simulated An increase in corn ethanol production from its 2004 level (3.41 BG) to 15 BG, off of the 2004 database. An increase in production and consumption of Bio- Gasoline produced from corn stover by 6 BG (or 9.0 BG ethanol equivalent), off of 15 BG corn ethanol. An increase in production and consumption of Bio- Gasoline produced from miscanthus by 4.7 BG (or 7 BG ethanol equivalent), off of 15 BG corn ethanol. An increase in production and consumption of Bio- Gasoline produced from switchgrass by 4.7 BG (or 7 BG ethanol equivalent) on top of 15 BG corn ethanol, 14
Scenarios Simulated Increase in the production and consumption of ethanol from corn stover by 9 BG, on top of 15 BG corn ethanol Increase in the production and consumption and consumption of ethanol from miscanthus by 7 BG on top of 15 BG of corn ethanol. Increase in the production and consumption of ethanol from switchgrass by 7 BG on top of 15 BG of corn ethanol. 15
Preliminary Land Use Changes (a) (b) (c) (d) 15 BG ETH Off of 2004 6 BG Stover Bio-Gasoline 4.7 BG Miscanthus Bio-Gasoline 4.7 BG Switchgrass Bio-Gasoline Land cover US EU Brazil Others Total Forest -331-80 42 144-226 Crop 971 126 82 899 2,078 Pasture -639-46 -123-1,043-1,852 Land cover US EU Brazil Others Total Forest 8 2 0 47 56 Crop -13-2 -2-15 -32 Pasture 5 0 2-32 -24 Land cover US EU Brazil Others Total Forest -153-16 8 24-137 Crop 106 25 15 173 319 Pasture 47-9 -23-197 -183 Land cover US EU Brazil Others Total Forest -550-45 20-16 -590 Crop 223 65 40 447 775 Pasture 327-20 -60-431 -185 16
Preliminary Land Use Changes - Ethanol (e) (f) (g) 9 BG Stover Ethanol 7 BG Miscanthus Ethanol 7 BG Switchgrass Ethanol Land cover US EU Brazil Others Total Forest 19 3 0 52 74 Crop -13-4 -3-25 -44 Pasture -6 1 3-28 -30 Land cover US EU Brazil Others Total Forest -221-21 11 26-205 Crop 134 32 20 222 408 Pasture 88-11 -31-249 -202 Land cover US EU Brazil Others Total Forest -784-61 28-29 -845 Crop 301 89 54 610 1,054 Pasture 483-28 -82-581 -208 17
(a) (b) (c) (d) (e) (f) (g) Biofuel Case Corn Ethanol Stover Bio-gasoline Miscanthus Bio-gasoline Switchgrass Bio-gasoline Stover Ethanol Miscanthus Ethanol Switchgrass Ethanol Biofuel Produced (billion gallon) Land use changes New Cropland Needed (1000 ha.) New Cropland Needed (ha./1000 gallons of biofuel) New Cropland Needed (ha./1000 gallons of ethanol eq.) 11.59 2078 0.18 0.18 6-32 -0.005-0.004 4.7 319 0.07 0.05 4.7 775 0.16 0.11 9-44 -0.005-0.005 7 408 0.06 0.06 7 1054 0.15 0.15
Biofuels Covered US Corn ethanol US soybean biodiesel Brazilian ethanol 19
Sensitivity Analyses Sensitivity of land cover changes with respect to changes in the food demand induced by higher food prices due to biofuel production. Sensitivity of land cover changes with respect to yield-to-price elasticity. Sensitivity of land cover changes with respect to cropland transformation elasticity. Sensitivity of land cover changes with respect to endogenous productivity change for cropland pasture. 20
Model Modifications Updated energy elasticities, Improved treatment of DDGS and oilseed meals and oils, Separation of soybean from other oilseeds, Separation of soybean oil from other vegetable oils and fats, Separation of soybean biodiesel from other types of biodiesel. Modified model structure for livestock sector. 21
Model Modifications Revised land conversion factor for new cropland, Incorporate cropland pasture for US and Brazil and CRP for US, Endogenous yield adjustment for cropland pasture, Greater flexibility in cropland switching. Substitution among soybean oil and other vegetable oils and fats 22
Household Demand Structure Household Demand for Private Goods CDE Energy Composite Non-Energy Commodities CES s ELEGY Coal Oil Gas Electricity Petroleum products & Biofuels Composite s ELHBIOIL Petro products (Oil_Pcts) Biofuels s biofuel Grain based Ethanol Sugarcane Ethanol Soy biodiesel Other Biodiesel 23
Firms Input Demand Structure Resources Value added & Energy Labor Land AEZ1 Crude oil s ESUBVA s ESAEZ AEZ18 Capital CES Firm s output s Non-energy intermediate inputs Capital-Energy composite Electric Non-coal s ELKE Composite Energy good s ELEN Non-electric s ELNCOAL Domestic s ELNEL Coal s ESUBD Imported s ESUBM Region 1 Region r Natural gas Petroleum products & Biofuels composite s ELBIOOIL Petro products (Oil_Pcts) Biofuels 24 Grain based Ethanol Sugarcane Ethanol Soy biodiesel Other Biodiesel
25 Nested Demand for Livestock Feed Feed Composite EFED=0.9 Livestock CROPS Processed Feed Energy-Protein LVFD=1.5 CRFD=1.5 OBCD=0.3 Sugar Crops Other Agriculture Other Grains Oilseed-Meal DDGS-Coarse Grains Intermediate inputs from livestock and processed livestock S-M-Soy ODBD=20 S-M-Oth CDDG=20, 25, and 30 DDGS Coarse Grains OBDS=10 OBDO=10 Soybeans Soybean Meal Other Oilseeds Other Meals
Land Cover and Use Nesting Land Cover 1 =-0.2 Forest Pasture Cropland Pasture-Land 3 =2.0 (1) Cropland-Pasture CRP 2 =-0.75 Crop 1 Crop N (1) In this land supply tree 1 and 2 are transformation elasticities and 3 is the elasticity of substitution between pasture land and cropland pasture in the livestock industry 26
Endogenous Cropland Pasture Yield Change We received comments on our previous work suggesting that the increased use of land for biofuels would lead to investments in increased productivity as land rents increased. This led us to introduce an endogenous change in cropland pasture productivity as cropland pasture rent increases due to higher demand for the resource. This change in productivity is a function of the change in rent and a new elasticity parameter. 27
Endogenous Cropland Pasture Yield Change af pasture : Cropland pasture augmenting technical change, A: Area under dedicated energy crop (0 in this analysis), B: Area remaining in cropland pasture, pf : Percent change in the cropland pasture rent, α: Scalar yield elasticity (0.4), β: Scalar yield adjustment factor (0 in this analysis), The yield-to-price elasticity is set to zero for cropland pasture. 28
New Database Modifications Split harvested area and production of soybeans from other oilseeds, The osd sector is divided into two industries of Soybeans and Other_Oilseeds, The vol industry divided into two industries of Vol_Soy and Vol_Oth which prodcue: Soybean oil and Soybean meal, Other vegetable oils and non-soybean meals We incorporated two biodiesel industries of Biod_Soy and Biod_Oth. 29
Land Use Change Results (ha/1000 gal. biofuel) Biofuel CARB 2009 Purdue 2010 Current Results Results with CP US corn ethanol 0.29 0.13 0.22 0.18 0.31 US soy biodiesel 0.63 0.94 a 0.33 0.64 Brazilian sugarcane 0.55-0.16 0.39 Complete details on land use change have been provided to CARB. 30
0-500,000-1,000,000 ha -1,500,000-2,000,000-2,500,000-3,000,000 Food consumpion is sensitive to price changes Food Consumption Sensitivity Corn ethanol Scenario Food consumption is Food consumption is fixed in developing fixed globally countries Forests Pasture -400,000 10% 18% -450,000 21% 48% ha 0-50,000-100,000-150,000-200,000-250,000-300,000-350,000 Soy biodiesel Food consumpion is sensitive to price changes Scenario Food consumption is Food consumption is fixed in developing fixed globally countries Forests Pasture ha Sugarcane ethanol 0-100,000-200,000-300,000-400,000-500,000-600,000 Food consumpion is sensitive to price changes Scenario Food consumption is Food consumption is fixed in developing fixed globally countries 16% 22% Forests Pasture The food consumption sensitivity results indicate that the land cover change is somewhat sensitive to changes in the food consumption assumption. 31
Yield-to-price Elasticity Sensitivity Corn ethanol Soy biodiesel magnitude of yield-to-price elasticity magnitude of yield-to-price elasticity 0-500,000 0.25 0.1 0.05 0-50,000-100,000 0.25 0.01 0.05-1,000,000 ha -1,500,000-2,000,000 forests pasture ha -150,000-200,000-250,000-300,000 forests pasture -2,500,000-350,000 ha -3,000,000-3,500,000 Sugarcane ethanol 0-100,000-200,000-300,000-400,000-500,000-600,000-700,000-400,000-450,000 34% 51% 36% 55% magnitude of yield-to-price elasticity 0.25 0.01 0.05 28% 42% forests pasture The results in all cases are sensitive to the value of the price-yield elasticity. Of the three, sugarcane is least sensitive, and soybean is the most sensitive. 32
Sensitivity on cropland transformation elasticity and cropland pasture endogenous technical change elasticity US corn ethanol US soy biodiesel Brazilian Sugarcane ethanol Biofuel Case Transformation Elasticity = -0.75 Transformation Elasticity = -0.5 forest Cropland pasture forest Cropland Pasture Area -290,330 2,118,901-1,828,303-246,078 2,237,912-1,991,761 ha/1000 gall -0.03 0.18-0.16-0.02 0.19-0.17 Area -18,286 267,915-249,803-9,639 272,147-262,557 ha/1000 gall -0.02 0.33-0.31-0.01 0.34-0.32 Area -95,315 465,295-369,887-36,457 543,158-506,862 ha/1000 gall -0.03 0.16-0.12-0.01 0.18-0.17 Biofuel Case US=0.4 and Brazil=0.2 US=0.0 and Brazil=0.0 forest Cropland pasture forest Cropland Pasture US corn ethanol US soy biodiesel Brazilian Sugarcane ethanol Area -290,330 2,118,901-1,828,303-550,067 2,011,577-1,461,333 ha/1000 gall -0.03 0.18-0.16-0.05 0.17-0.13 Area -18,286 267,915-249,803-62,022 247,766-185,742 ha/1000 gall -0.02 0.33-0.31-0.08 0.31-0.23 Area -95,315 465,295-369,887-186,249 449,784-263,459 ha/1000 gall -0.03 0.16-0.12-0.06 0.15-0.09 33
Thank you! Questions and Comments For more information: http://www.ces.purdue.edu/bioenergy http://www.agecon.purdue.edu/directory/d etails.asp?username=wtyner